Can we make wise decisions to modify ourselves?
Some problems and considerations
University of Manitoba
Journal of Evolution and Technology - Vol. 29 Issue 1 – May 2019 - pgs 1-18
Much of the human enhancement literature focuses on the ethical, social, and political challenges we are likely to face in the future. I will focus instead on whether we can make decisions to modify ourselves that are known to be likely to satisfy our preferences. It seems plausible to suppose that, if a subject is deciding whether to select a reasonably safe and morally unproblematic enhancement, the decision will be an easy one. The subject will simply figure out her preferences and decide accordingly. The problem, however, is that there is substantial evidence that we are not very good at predicting what will satisfy our preferences. This is a general problem that applies to many different types of decisions, but I argue that there are additional complications when it comes to making decisions about enhancing ourselves. These arise not only for people interested in selecting enhancements but also for people who choose to abstain.
Alexis Madrigal, an American journalist, decided to implant a magnet under the skin of his fingertip. He came to this decision while interviewing bodyhackers for an article he was researching. The bodyhackers spoke with such enthusiasm about how having magnets under their fingertips opened up new sensory worlds that Madrigal decided to try it for himself. Once the novelty wore off, he discovered that he did not like having a magnet under his fingertip all that much. He reports that he sometimes worried about the possibility of infection, and that he would likely get the magnet removed at some point (Madrigal 2016).
The decision whether or not to install a magnet under the skin seems a minor one, akin to getting ears pierced. There are, however, more substantial interventions we can undertake to change our abilities or our bodies (e.g., Neil Harbisson implanted a device in the bone of his skull that allows him to hear colors (Jeffries 2014)), and it is likely that more profound decisions will face us in the future. Many authors discuss the safety of future possible enhancements (Chatwin et al. 2017; Torres 2016; Douglas 2015; Bostrom and Sandberg 2009), their influence on our society (Crittenden 2002; Fukuyama 2003), and their potential influence on our moral selves (Archer 2016; Agar 2014; Persson and Savulescu 2013; Sandel 2009) and our sense of self (Edelman 2018, Coeckelbergh 2011, Cabrera 2011, Sandel 2009). In what follows, I will argue that, even if we can successfully address these issues (a tall order in some cases, but reasonably easily met in the case of Alexis Madrigal’s choice), we still face the problem that we cannot always know whether our decision to modify or not modify ourselves will satisfy our preferences.1 Madrigal discovered only after the fact that he preferred the sensation of an unmodified finger to one with a magnet installed. Furthermore, this problem faces not only those who are interested in bodyhacking and enhancement, but everyone, including the abstainers. Once a modification is contemplated and considered, abstaining is still a choice made without sufficient information to know that it is the best route to satisfying our preferences.
The decision problem I am identifying is known in the decision theory literature as making a decision under uncertainty. After a few preliminaries, I will give a brief, non-technical description of the problem of deciding under uncertainty. Then I will present five routes for addressing uncertainty. The bulk of this article will examine these five routes and the ways in which they can and cannot be used to make wise decisions about enhancements and body modifications (I use the term “modification” to signal that in some cases we might want to change ourselves without making ourselves “better” in some sense). While it is true that we normally have difficulties figuring out which choices will satisfy our preferences, I will consider the specific problems that face us when we decide whether to adopt or reject getting modifications.
Toward the end, I will consider whether the decision difficulties for modifications are on a par with other momentous decisions we face. For example, people routinely decide whether to get married, or to have children, or to pursue a certain career, and we make all these decisions without sufficient information to know with certainty that we have made the right choice. I will argue that, for certain types of enhancement or modification decisions, the lack of information is a more serious problem than for these other momentous decisions. To clarify, the thesis is not that we cannot make rational decisions about modifying ourselves. We can, if we define “rational decisions” as decisions that follow some reasonable set of decision-theoretic rules. The thesis is also not that we cannot make the right decision, if this means the one that leads to the best outcome. It is, after all, possible even for foolhardy decisions to lead to lucky outcomes. Instead, the thesis is that, when we are dealing with an information gap, the extent to which rational decisions track right decisions weakens, and the information gap for certain modification decisions is larger than for other types of momentous decisions.
My focus here is on decisions we make on behalf of ourselves, as opposed to decisions we might make on behalf of others. I will also concentrate on decisions about actions that are neither morally required nor forbidden. Installing magnets under the skin seems to fall under this category, and is a relatively minor decision. Note, however, that non-obligatory, permissible actions can be quite profound in their consequences or implications. For example, in 2006 the Guardian reported on a case where a man’s penis was damaged beyond repair in an accident. He opted for a penis implant, which seems both non-obligatory and permissible. The implant was medically successful. Nonetheless, surgeons removed the transplanted penis two weeks later because the recipient and his wife could not psychologically accept it (Sample 2006). I think it is easy to imagine why the man consented to the transplant surgery, and easy to imagine why the results would be sufficiently psychologically disturbing to warrant the removal of the transplant. It would also be understandable if a year or two down the road the man regretted his decision to remove the transplanted organ. This case points to the difficulties in making decisions about how we might modify our own bodies.
When decisions made on behalf of oneself are about permissible, non-obligatory actions, many of the reasons we have for them will be based on whether the decision is likely to satisfy our personal preferences. These types of reasons are subjective, and they vary from individual to individual. I will follow L.A. Paul (2014, ch. 2) in calling these types of reasons first-personal. While it is true that, for some decisions, more objective, or other-directed, reasons might prevail (for example, we might prefer to remain unmodified, but might decide to improve our intelligence in order to increase the odds of discovering a medical treatment that would save the lives of many), a wide variety of decisions might reasonably involve mostly first-personal reasons. I will focus on decisions based on first-personal reasons in this article, although there are further interesting conversations to be had about decisions involving a mix of first-personal and, in the relevant sense, objective reasons (Martens 2016 briefly touches on this subject, although not in the context of decision theory).
Paul argues that, when the reasons for making a decision are first-personal, what we need to know is the “what it is like for me,” or the “what it is like for the agent” (the WIILFA) (Paul 2015b, 808–809). The WIILFA has two components. The “what it is like” refers to lived experiences, a kind of “thisness” of the now. The “what it is like” cannot be fully communicated from one person to another. For example, there is a difference between being told, “You will hear an extremely unpleasantly loud noise,” and actually experiencing that noise (Hsee, Hastie, and Chen 2008, 233). The “for me” or “for the agent” part signals that what matters in making a first-personal decision is whether the agent will prefer the outcome, rather than whether people in general, or even people similar to the agent, will prefer the outcome. Some decisions are going to be more WIILFA-dependent than others. For example, the WIILFA of “it feels great to be smarter” seems less important than the WIILFA of “it feels great to be able to hear colors.”
2. Rational decisions
The standard procedure for making a rational decision involves making a calculation based on the following pieces of information:
1) The possible ways the world could be (the state) that are relevant to each choice.
2) The probabilities that the possible states will occur should the agent make that choice.
3) The outcomes based on the states.
4) The expected values of the outcomes.
The decider will often lack full information about the above four pieces of information. Furthermore, lacking information about the first will influence the other three. Before we get to problems about lacking information, let’s consider a relatively simple example about buying a lottery ticket (Table 1). The choices are on the left-hand side of the table. They are: buy ticket with number x; do not buy ticket. The states are in the top row of the table. They are: Ticket number x wins; Ticket number x does not win. The outcomes are the results of the choice and the states, and are numbered in the table from best expected value to worst expected value.
If we do not consider the probabilities of each state occurring, then buying the ticket seems the rational choice because it can give us the highest value outcome. Obviously, then, we should consider the probabilities. While the expected value of winning a lottery is high, the probability of winning is low, and so we are more likely to end up with the worst outcome (outcome 4), rather than the best, if we buy the ticket. Unless we derive an additional value out of purchasing the lottery ticket (for example, coworkers might purchase lottery tickets together as a kind of community-building activity), we should not purchase it. (Outcome 2 can become the worst outcome if the decider knows she could have bought the winning ticket. Suppose, for example, she normally selects a specific sequence of numbers but opts out in the week that those numbers win. Then the decider gets to add “deep regret” to outcome 2.)
Now let us consider whether I should get a magnet implanted under my fingertip (see Table 2).
This is a simplified table. For example, I have excluded the possible state where my finger gets infected. I have also excluded a category for states that we have not anticipated. I want to note two things. First, the states in question are experiential and subjective (whether I like the sensations). Second, since I have never had an implant, the decision is about whether to have a novel experience. The novelty and subjectiveness contribute to my not knowing the probability of whether I am the kind of person who will like the sensations or the kind of person who will not. This implant decision problem is different from the lottery case. In the lottery case, we do not know in advance which state will occur (winning or not winning), but we do know that the probability of winning is very low. We know the probabilities of the states in the lottery case, but not in the implant case (or, we might subjectively assign probabilities with a low level of confidence in those probabilities, but I will not get into such technical issues here). This lack of knowledge will impair our ability to choose the act most likely to lead to the best outcome.
The problem of making decisions without knowing with confidence the probabilities of the states is well known in the decision theory literature, with many approaches offered (for example, Wald 1950, 18; Hurwicz 1951; Savage 1972, ch. 9; Skyrms 1990, 112–14 ; Weirich 2004, ch. 4; Stoye 2011; Buchak 2013, ch. 1). For example, one could follow the rule (the maximin rule) where one makes choices to avoid the worst outcome, which in this case is outcome 3 (get the implant + do not like the implant = new unpleasant sensations). Then I should not get the implant. A different rule would have me considering whether I am risk-averse or risk-adventurous, and then make my selection accordingly. Since I am risk-averse, I would make the choice to avoid the worst outcome (outcome 3). Alexis Madrigal seems more adventurous, at least from his telling of the story, and he justified his choice because he had an interesting learning experience. According to this rule, the decision to get the implant is irrational for me, but rational for Madrigal given our different attitudes toward risk (and given that the risks associated with implanting magnets are quite low). Notice that these decision rules are to be applied at the time of making the decision. Below I will discuss strategies for gathering information prior to the time of making the decision.
If we merely apply some reasonable decision rule when operating under an extreme lack of information about the probabilities of the states, then we have reasons for our decisions. Making a reason-based decision is not the same thing as making the right decision. The right decision is the one that actually brings about the best outcome. When operating under an extreme lack of information, a wedge is driven between reason-based decisions and right decisions. In this context, reason-based decisions might not perform much better than a guess. Notice that this problem is symmetrical in the following way. A rule that prioritizes the status quo is no more likely to track the best outcome than a rule that prioritizes exploration. When facing modification decisions, we are no more likely to get it right if we reject the modification than if we adopt it. It is possible that if we could experience the WIILFA of the modification in advance of making the decision, we might conclude that the status quo is entirely unacceptable. The reverse is also possible. Neither transhumanists nor bioconservatives have an edge here.
Not all is lost. Sometimes we can take steps to reduce the amount of uncertainty. Here I am concerned with applied rather than theoretical decision theory, so the goal is to collect strategies that can be used by real agents rather than ideal agents. Several authors (Burnett and Evans 2016; Krishnamurthy 2015; Dougherty, Horowitz, and Sliwa 2015; Pettigrew 2015; Paul 2014, 2015a, 2015b, 2015c; Weirich 2004) focus on realistically usable strategies. With the exception of the first strategy surveyed below, these differ from rules like the maximin rule in that they concern how we should gather information or frame the decision. These are strategies for preparing for a decision prior to applying a decision rule. I will survey these five strategies before applying them to decisions about body modification. Not all of these strategies will turn out to be good strategies for making decisions about body modification, and all of them will have limited applications.
3. Five strategies for dealing with uncertainty
3.1 – High need or low cost
The idea here is that, while we face a lack of information about the WIILFA of the novel experience, we may still have enough information to make a choice. In the high need case, the costs of not changing are high enough to motivate trying a novel experience. For example, having a currently untreatable fatal disease could make it rational to try an experimental treatment despite a lack of information about effectiveness or safety (for an introduction to dominance, see Whitmore and Findlay 1978, 24–27). In the low-cost case, while we might not know whether we will prefer the WIILFA of the novel experience, the costs of trying are low enough to be worth finding out. The low-cost case can be buttressed by a decision rule focused on avoiding regret. The strategy of appealing to high needs or low costs clearly has limited applications, as will be discussed in section 4.
3.2 – Curiosity
We could justify trying the novel experience on the grounds that we value discovering what it will be like. This way, at least one of our preferences – the preference to discover what it will be like – will be satisfied regardless of what the other aspects of the experience will be like. This is a reframing strategy proposed by Paul in response to the failures associated with the next option (Paul 2014, ch. 4). As we shall see in section 4, the curiosity option has extremely limited applications. Furthermore, the adventurous might find themselves regretting their adventures if the outcomes are bad enough.2
3.3 – Imagination
We could use our imaginations to try to determine what the novel experience will be like for us. Paul refers to this as a natural approach, which we use for many types of decisions (Paul 2015c). For example, when deciding on purchasing a new home, one might begin the decision process by imagining living in the home.
Paul also insists that using our imagination is a necessary approach when decisions depend on the value of the WIILFA for our future selves undergoing the experience. Imagination, Paul argues, is the only way to grasp what our unique future selves will be like because empirical studies can only be about other people who may or may not resemble our future selves (2015b, 448–87). The problem, which Paul acknowledges and will be elaborated on in section 4, is that our imagined future self is not likely to resemble our actual future self in the way that we need.
3.4 – Prototyping
Weirich (2004, 27) points out that rational agents will not simply rest content with applying decision rules to the information they have, but will collect new information whenever possible (a sequential versus a static approach to decision theory). He highlights the strategic value of making a series of decisions, each of which leads to acquiring new information that influences the next decision. In their self-help book on making decisions, Burnett and Evans offer similar advice from a less academic perspective (2016, ch. 6). They recommend prototyping, which involves seeking out lower risk experiences that are similar to the novel experience we are considering trying. For example, if we want an RFID chip implant in our hands that will allow us keyless entry into our homes, we could try the chip on a wearable device first. Krishnamurthy (2015) offers a similar suggestion when she argues that we can obtain information about what it will be like to be a parent by spending lots of time caring for children. This involves an argument by analogy, that the prototyped experiences we have had in the past are sufficiently similar to the experiences we will have in the future if we make the higher risk decision. This strategy and the next are the most promising routes, although their limitations will be discussed in the next section.
3.5 – Collecting empirical data
The most common advice in the literature is to collect more information about the probability of the desired states by consulting empirical studies (quantitative or qualitative) of others who have already made the same decision (Krishnamurthy 2015; Dougherty, Horowitz, and Sliwa 2015; Pettigrew 2015). If we are considering a body modification, we should find out if others who have already tried the modification found it to be a valuable experience. This also involves an argument by analogy, to the extent that we are relying on those people being similar to us in relevant ways. This is not a watertight argument, however. After all, they might prefer the experience while we do not.
Note that strategies 3, 4, and 5 (imagination, prototyping, and collecting empirical data) involve trying to gather information on the probabilities of whether the subjective experiences will be valuable to us. In other words, the problem of uncertainty is addressed by reducing the level of uncertainty. Strategy 2 (curiosity) does as well, but in a different way. It involves shifting focus from future states including or not including the WIILFA to states about discovering or not discovering the WIILFA. The agent might not know whether she will like the WIILFA, but she will have information about whether she values discovery or is risk-averse.
Strategy 1 (high need or low cost) does not involve basing justification on an approximation of the expected value of the WIILFA of the novel experience at all, but on either the known bad expected value of the WIILFA of the known outcome or the known low risks of the novel option. Whatever might be the expected value of the novel WIILFA, it stands a good chance of being better than the WIILFA of the known outcome, or in any event it is unlikely to be bad enough to avoid.
4. The main argument
It is now possible to state an outline of the argument of this article:
1) In order to be reasonably confident that our decisions will satisfy our first-personal preferences about the WIILFA, we need to reduce sufficiently the amount of uncertainty associated with a decision. This applies equally to the decision to stay the same as it does to the decision to change or enhance oneself.
2) For many people, and for certain types of modifications, the above five strategies are either inadequate or unavailable for reducing sufficiently the amount of uncertainty associated with a decision. (This is especially likely in the early stages of modification research. If modifications come into widespread use, we will be presented with different problems, discussed below.)
3. Therefore, if the above five strategies are the only routes for sufficiently reducing uncertainty, then for many people, and for certain types of modifications, a reasonable amount of confidence that decisions will satisfy first-personal preferences cannot be obtained.
It is worth noting that we are currently facing this problem. I have already not installed magnets under my fingertips many times. If you are like me, then you will also have made many such decisions without a great deal of thought.
One obvious vulnerability of the argument, implicit in the wording of its conclusion, is the possibility of some strategy not in the list of five mentioned above. Any further justification strategies will make for interesting discussion at another time. I cannot deal exhaustively with the issue on this occasion, although at the end of the article I will speculate briefly on routes we might take. Meanwhile, the bulk of what follows will focus on justifying premise 2. I will examine each of the suggested strategies, in order to show why they do not always apply.
4.1 – High need or low cost
It is possible to make rational decisions when the outcomes of a novel course of action are unknown if the outcomes of the other course of action are known to be sufficiently worse. For example, in the 1980s, many people infected with the human immunodeficiency virus (HIV) were willing to try experimental drugs because the known outcomes of not trying the drugs were severe.
It would be far too quick to rule out this justificatory strategy for enhancements on the grounds that enhancements are not treatments. Aside from the oft-noted blurry boundary between enhancements and treatments (see, for example, Agar 2014; Lara 2017), sometimes enhancements are still solutions to problems. This justificatory strategy is not limited to life and death situations. All that is needed is enough information to be able to conclude reasonably that, whatever the outcomes might be from the novel course of action, they are very likely to be better than those of the non-novel course.
I have grouped the high-need reason together with the low-cost reason because they both depend on a proportional bar of risk. If the possible risks of the novel action are reasonably low, then the problem it is intended to solve does not need to be severe. Consider the magnet case. While it is novel, and we might be in the dark about whether we will like the insertion of a subcutaneous magnet, we still have a pretty good idea of some of the risks. They are fairly low. We know, to use an absurd example, that installing magnets under our skins will not turn us into frogs. Since the range of possible outcomes of installing a magnet are not extremely bad, the problems that magnets could solve also do not need to be extreme. They only need to outweigh the range of possible negative outcomes. Let us consider a few examples of enhancements that are solutions to problems.
Amal Graafstra had the problem of leaving his keys in his office and locking himself out. He solved his problem by installing an RFID chip implant under the skin of his hand that unlocks his door. Now he does not lose his keys anymore (Graafstra 2013).
Neil Harbisson is a visual artist who was born completely color blind. Nonetheless, he wanted to work artistically with colors, so he installed a device that is fused to his skull and allows him to hear colors (Jeffries 2014).
In an article on bodyhacking published in the New York Times, Hylyx Hyx is quoted as saying, “I’m used to having weird feelings about my body . . . I use ‘they’ pronouns. I don’t care about most of my meat, so this is a way to have control over a part that I chose” (Hines 2018). One way to interpret Hyx is as expressing a lack of connection with their body, and a desire to control it. Through body modification, Hyx can now control and connect to their body.
Hylyx Hyx also stated that they are a “submissive for science” (Hines 2018). One interpretation of this claim is that Hyx is willing to be an experimental subject. It is possible that such experiments could lead to treatments that address not the high needs that Hyx has, but needs that others have. This, however, is not a common motivation.
Liao, Sandberg, and Roache (2012) argue that if we enhance ourselves and our offspring so that we consume fewer resources and are more intelligent and ethical, we might be able to find a solution to the pressing problem of climate change.
Back to the HIV case. It would not normally be rational to try an untested drug if another known effective drug were available. In the above four examples, there are alternative solutions to the problems. Graafstra and Harbisson could use wearables. Hyx could stick to piercings, tattoos, or diet plus the gym to control their body. For the problem that Liao, Sandberg, and Roache raise, our respective governments could better incentivize good environmental behaviors on our part (which seems a more likely successful route than incentivizing that we modify ourselves and our offspring).
It is possible that Graafstra, Harbisson, and Hyx would find my alternative proposals unacceptable. They could, for example, have deep identity-based reasons for why implants, even implants with potentially negative consequences, are a more palatable way to go than the alternatives. I do not know them personally, so I cannot say what their reasons are, but it is certainly conceivable, and even likely, that some people might have deep identity-based reasons for wanting to modify their bodies in certain sorts of ways, even if Graastra, Harbisson, and Hyx themselves do not. For an identity-based reason to count as a high enough need to swamp other considerations, it would need to be similar to what many trans*3 people experience when deciding whether or not to transition. Rachel McKinnon offers a version of this decision-making strategy when she points out that, while trans* people might not know what it will be like to transition, they are often faced with dire situations if they do not. To support this, she cites studies of high suicide rates of trans* people (McKinnon 2015, 423–24). Similarly, White cites studies of the suffering of people with Body Integrity Identity Disorder who are prevented from obtaining elective amputation for the purposes of aligning their bodies with their identities (White 2014, 226).
Even if, however, we can successfully argue that identity expression through body modification is important enough to swamp the possible negative outcomes, it also seems likely that this method of justification will be available only to some people. Many others will not feel driven by identity reasons to expand their abilities or modify their bodies. Thus, to the extent that the identity threat argument works, it will work only for some. Furthermore, many of the enhancements and modifications that we might consider are not solutions to any problem at all. Therefore, the high need/low cost approach is a justificatory strategy with limited application.
4.2 – Curiosity
The adage, “Fools rush in where angels fear to tread” is one description of what happens when people base a decision to try the novel on curiosity. Here is another description. Paul proposes that we can reframe the decision problem in a way that permits a rational decision to be made. The idea is this. Instead of trying to ascertain which option will increase first-personal value based on an unknowable WIILFA, we might base the decision on whether we value finding out the WIILFA. For example, if we value knowing the WIILFA of having magnets under the skin of our fingertips, we will satisfy that value regardless of what the WIILFA turns out to be (Paul 2014, ch. 4). In Alexis Madrigal’s case, given the tone of his report, it seems that he valued having the implant experience for the purpose of finding out what it would be like, even though, in the event, he did not particularly care for the what-it-was-like. (We can use this reframing strategy not just for values such as a discovery preference, but also for values such as controlling one’s body.)
We can also use this reframing strategy for rejecting the novel experience. For example, it is quite reasonable to value not finding out the WIILFA of age-related cognitive decline, so taking a preventative drug to avoid having that discovery is rational.
There are at least two ways in which the valuing discovery option is limited. First, for certain types of decisions, it counts as a terrible reason. Consider Paul’s stock example of choosing to have a child for the first time. Having a child for the sake of wanting to find out the WIILFA seems incredibly flippant. This level of flippancy might be reasonable when it comes to more trivial decisions like installing magnets or RFID chips under the skin, but not for deciding to have a child or deciding to take on a more profound enhancement. The level of risk matters here. Also, in the case of deciding to have a child, other parties are deeply affected by our decision. (Interestingly, the flippancy seems to disappear when we consider the decision to not enhance or to not have a child. Wanting to not find out the WIILFA seems acceptable here.)
Second, only a relatively small number of people will be able to make use of the curiosity-justification strategy. Basing a decision on valuing discovery is available to those for whom wanting to find out the WIILFA is a sufficient reason to adopt an enhancement. It is also available to those for whom wanting to not find out the WIILFA is a sufficient reason to avoid an enhancement. Basing a decision on valuing discovery is not available for the following two groups: those for whom these wants are not sufficient (e.g., I am kind of curious about having an RFID chip implant, but not enough to override my concern about uncertainties about the WIILFA); and those who do not have wants either way about discovering the WIILFA. Let us explore the latter point.
There is a distinction between wanting to not find out the WIILFA, and not wanting to find out the WIILFA. The first involves an active desire to avoid learning about the experience. For example, I have an active desire to avoid finding out what it feels like to jam a pencil into my hand. The second involves the absence of an active desire. For example, I do not have any desires one way or the other to find out the WIILFA of having arms that are 1 centimeter longer than they currently are.
While it seems pretty clear that the bodyhacking movement is driven forward in part by curiosity (along with other desires, such as controlling one’s body) – and so, many bodyhackers can avail themselves of this justificatory strategy – the bodyhacking community is relatively small. It seems very likely that many people will fall into the categories of either not having a sufficient desire to discover/avoid finding out the WIILFA or not having a desire at all about the discovery aspect of enhancement. These people, then, are still faced with the problem of justifying their decision either to try or to avoid the novel enhancement.
4.3 – Imagination
A large body of literature in psychology shows that we are terrible at predicting whether, or how much, we will prefer an experience (Wilson and Gilbert 2005; Gilbert et al. 2009; Walsh and Ayton 2009). We are terrible at this even when the predictions in question involve events we have experienced many times in the past. For example, if we are still full from the last delicious meal, we tend to underestimate how much we will enjoy the next delicious meal. Yet we have gone through the process of moving from satiety to hunger many times in our life. Given this, we can expect to be terrible at predictions about the first-personal values of novel experiences. As some authors have previously pointed out, future research on human enhancements might take us in directions that boggle our imagination (Mihailov and Dragomir 2018).
It is worth taking a few moments to review one of the identified errors we tend to make when trying to ascertain whether we will prefer a new experience, and whether the price we pay for that new experience is suitable. Forewarned is forearmed. We tend to overestimate how much we will react to a future event (Wilson and Gilbert 2005). To use Hsee, Hastie, and Chen’s (2008) example, when we first move into a larger home, our initial feelings of pleasure will be stronger at the start. Later, however, we will adjust to this larger home, and the intensity of the pleasure will decrease. Hsee, Hastie, and Chen refer to this as the distinction between acquisition and consumption. When deciding whether or not the larger home is worth the greater expense, we often make the mistake of basing our decision on the prediction of the acquisition experience rather than the consumption experience.
Similarly, when considering modifying the body, we might make the mistake of basing our assessment on a prediction of the acquisition experience rather than the consumption experience.
4.4 – Prototyping
If the problem of uncertainty arises because we do not have past experiences that are sufficiently similar to the future experiences we are considering, we may make lower risk decisions that give us information about those future experiences. Earlier, I gave examples such as wearing an RFID chip on a ring before implanting it in the hand. For another example, Neil Harbisson, prior to fusing a device to his skull that allows him to hear colors, wore a prototype that strapped onto his head. This allowed him to gather significant information on whether or not the device worked, and whether or not he could learn the color-to-sound language (Stix 2016).
Prototyping has its limits. One feature Harbisson could not prototype was the WIILFA of having the device fused to his skull. Another feature that cannot be prototyped is the long-term WIILFA. As previously mentioned, there is a difference between the acquisition experience and the consumption experience, and this difference matters.
Other enhancement choices might not be suitable for prototyping at all. For example, it is difficult to imagine how one would prototype being substantially smarter (a little bit smarter, yes, but substantially smarter is more difficult).
4.5 – Collecting empirical data
When making decisions, it can be useful to find out how other people value the outcomes of their choices. We could talk directly to people about their experiences, or read the narratives contained in qualitative studies, or look at the results of quantitative studies. For this approach to provide us with information about the probabilities of outcomes that satisfy our preferences, two conditions must be met. First, there must be people who have already made the decision in the past. Second, those people must be sufficiently similar to us in relevant respects for the information about their experiences to be predictive of ours. Relevant dissimilarities will weaken the argument by analogy.
Let us consider three stages. Stage 1 is where nobody has yet tried the modification. Stage 2 is where only early adopters have tried it. Stage 3 is where the modification is widely used. Obviously, the stages will not, in practice, be sharply delineated. Nonetheless, the question remains. How do we get from Stage 1 to Stage 2, and from Stage 2 to Stage 3?
When moving from Stage 1 to Stage 2, the first condition cannot be met. We cannot find out how others have valued the experience because nobody, to date, has had the experience. Early adopters will need to make use of some other method of justification. To quickly review, the high risk or low cost strategy requires either that the status quo be high risk or the modification be low cost. There will be some modification decisions that do not fall into either of these categories. Valuing discovery for its own sake might be appealing to early adopters, but again, the strength of the desire for discovery needs to increase proportionately with the risk of the modification. Imagination is very inaccurate, and pretty much a non-starter. Prototyping is useful, but only certain aspects of the novel experience can be prototyped.
What are we left with? Here are three ways to move from Stage 1 to Stage 2:
1) Some people might adopt early because their modifications are also treatments.
2) Those with a high desire for discovery might move us from Stage 1 to Stage 2.
3) Those who make decisions without giving them much thought might also show up in the early adoption crowd.
It is interesting to think that, in moving toward a posthuman future, we might be relying on thoughtless behavior. As a quick aside, however, thoughtless decision-making can sometimes be beneficial. I imagine that many can identify with the claim that, had we given enough thought to certain decisions (e.g., having a child or training to become a professor), we might not have made those decisions, but we are nonetheless glad that we did.
When we move from Stage 2 (early adopters only) to Stage 3 (widespread use), we now have people who can report on their experiences. The question is whether early adopters are sufficiently similar to late adopters to support an argument by analogy. Since we are focusing on decisions that are based on satisfying preferences about the WIILFA, what matters is whether early adopters are sufficiently psychologically similar to late adopters. Physiological similarity between the two groups is important for establishing medical safety and effectiveness, but psychological similarity is needed to support the claim that, because the modification satisfied the preferences of the early adopters, it is likely to satisfy the preferences of the late adopters. I will argue that there are roadblocks to establishing similarity in preference sets between the two groups. These are not insurmountable, but they are cause for concern. We will begin with general problems on reports of the WIILFA, and then move to problems more specific to modifications.
The first general problem has to do with interpreting other people’s reports of their WIILFA. Hsee, Hastie, and Chen observe that, while we can use other people’s reports to provide information, information is “cold” while experiences are “hot.” There is an important experiential element that does not get communicated (Hsee, Hastie, and Chen 2008, 233).
The second general problem is with how people report on their decisions. People tend to rationalize decisions after the fact for a variety of complicated psychological reasons (Mather and Johnson 2000; Mather, Knight and McCaffrey 2005; Stoll Benney and Henkel 2006). This reduces the reliability of testimonials. On the plus side, the tendency to rationalize makes it likely that, whatever future decisions we make, we are likely to rate them more positively than, perhaps, we should.
For problems more specific to modification decisions, the reasons early adopters have for their decision might be different from those of late adopters, and these differing reasons might reflect different preference sets. Consider the early adopter who selects the modification as a treatment. One significant preference is to be relieved of the condition being treated. For the early adopter who selects the modification out of a sense of adventure, preferences for discovery, and even risk, will be salient. The thoughtless early adopter might have a positive attitude toward risk, at least with respect to the selected modification (although there can be many reasons for thoughtlessness). In sum, the early adopters might include some people who are merely thoughtless, but others are likely to have relevantly different preferences from the late adopters. One question that remains is whether these different preferences influence earlier and later groups’ perceptions of the WIILFA.
There is data to suggest that the context in which a decision is made influences the decider’s perception of the experience and the extent to which it satisfies preferences. For example, if a Positive Experience 2 follows another Positive Experience 1, we are inclined to rate Positive Experience 2 less highly than we would if it had followed Negative Experience 1 (Kahneman 1992). For someone selecting a modification as a treatment, and assuming the treatment is successful, we have a positive experience following a negative one. For another example, if a decision leads to the satisfaction of a goal, we tend to rank the experience more highly than if the decision leads to failing a goal, even if the actions and outcomes are the same (Heath, Larrick, and Wu 1999). For those motivated to modify out of a sense of adventure, this type of framing could influence reports of preference satisfaction.
It is striking that, given the identifiable differences in preferences between late adopters and those early adopters who are motivated either by needing a treatment or by a desire for discovery, the most suitable group for building an analogy between early and late adopters is the sub-set of early adopters who acted thoughtlessly. The thoughtless group is more likely to be diverse, since there can be many reasons for acting without thought. One issue that will matter is whether the thoughtlessness is global or local. If it is global, then it is hard to see how late adopters can trust the reports given by the thoughtless.
To be sure, the decider and the reporter (i.e. the early adopter) can keep these factors in mind. The reporter can be careful in how she articulates the nature of her experiences. The decider can sift through this information with awareness of the roles that psychological biases might play. The late adopter does not have perfect information, but the early adopter’s report gives him more information than he had in Stage 1. In assessing this, the late adopter needs to be aware that the differences between early adopters and late adopters weaken the analogy between their respective experiences. How much this matters depends on the magnitude of the risk if one makes the wrong decision. It also depends on just how different the late adopter is from the early adopter. Realistically, attitudes to early and late adoption will come in degrees. We might have early-early adopters who provide reasonable information for early adopters, and then the early adopters can provide reasonable information to early-late adopters. And so on. This would be a gradual rolling out of inferences. That said, in the early stages of Stage 2, the late-late adopters will still have a significant information gap. I emphasize again that this is a problem not just for those considering adopting a novel modification, but also for those who are not. Failing to make a decision and deciding to avoid an experience are still choices.
We also need to consider the problem in moving from Stage 2 to Stage 3 in the context of certain transhumanist philosophies. Neil Harbisson and Moon Ribas, co-founders of the Cyborg Foundation, articulate a vision of a world in which people design their own modifications in keeping with their specific desires about how they want to interact with the world (Harbisson and Ribas 2018). Clearly, on this vision, if we design a unique modification, we will not have access to information about how others have experienced it. I should note that Harbisson and Ribas recommend prototyping on their website.
5. Concluding remarks
In an article about the ethics of human enhancement, Norman Daniels recalls an old joke about a traveler asking for directions of a farmer. The farmer, after considering a variety of routes, says, “You can’t get there from here.” Daniels’ point is that we cannot ethically and safely get from our current world to one where modifications are profound and widely used (Daniels 2009, 38–41). My point is less emphatic. If we want to get from here to there while making informed decisions about whether the modifications will provide us with WIILFAs that satisfy our preferences, we will encounter significant roadblocks.
It might seem from the foregoing that I am recommending a bioconservative position. I am not, and these are some reasons why I am not. As already mentioned, neither the bioconservative nor the transhumanist has the edge on increasing the odds of making the right decision (the one that will yield the best WIILFA). Furthermore, if we think that we can avoid this decision problem by simply avoiding research into enhancement and modification techniques, then we are mistaken. Avoiding research is itself a potentially mistaken choice. Finally, and this is more of a confessional comment than anything else, I am personally enthusiastic about the possibility of a posthuman future such as envisaged by transhumanists. The problem is that the future I enthusiastically imagine is one with all the bugs and decisions already worked out, and I do not know how we can get from here to there. I applaud Hylyx Hyx for their willingness to be a “submissive for science” (Hines 2018), but I personally am not willing to go down that road.
I think that Ronald Dworkin put his finger on why the prospect of new human enhancement techniques is so troubling. Dworkin, in trying to sort out what the reasons might be for the “playing God” objection to human enhancement, observes that when scientific developments present us with choices we did not previously have, our views on ethics and justice get changed in a way that he calls “seriously dislocating” and requires a significant readjustment period (Dworkin 2002, 444). Dworkin argues that the reason our views on ethics and justice get changed is because our ideas about responsibility depend on the border between chance and choice. New scientific developments can change this border. Similarly, the prospects of novel enhancements and modifications present us with a “seriously dislocating” set of new decisions.
As Bostrom and Ord point out, modification decisions are not the only ones we have to make without sufficient information. Decisions about careers, becoming parents, and getting married are all high stakes decisions that we make without knowing whether they will turn out for the best (Bostrom and Ord 2006, 657). There is, however, one key difference between, for example, the decision to become a parent and the decision to adopt a radically new way of sensing the world. Many people like us have become parents before. We do not have a guarantee that our choice to become a parent will, on balance, be a good choice, but we do have access to a reasonable amount of information about this choice. Until we get to Stage 3 with modifications, decisions about modifying ourselves in novel ways will not be on a par with decisions about becoming parents or getting married or selecting a career.
Where do we go from here? Novel decisions are coming whether we like it or not. Marcus Arvan has an interesting answer, which he raises in the context of discussing transformative choices such as whether to become a parent. He proposes that instead of trying to fix the decision problem, we should work on becoming more resilient people so that we can adapt fruitfully to whatever outcomes arise (Arvan 2015). While this sounds like sage advice, I am troubled by the thought that, should I wind up making a modification that provides me with a negative WIILFA, I should rest content with the knowledge that I am the kind of person who will make the best of it. The bioconservative faces a similar problem in resting content with the idea that, even if he avoided selecting a beneficial modification, he should make the best of it. More disquieting is the tacit admission that, because of our psychology and our lack of access to information, we simply do not have a good decision toolkit available to navigate the types of decisions we will soon be facing about how we want to change our bodies and ourselves.
More promising are the suggestions that we attend to prototyping and empirical evidence. Here are just a few brief considerations. As already mentioned, not all aspects of a novel experience can be prototyped. One point of concern is that it will be particularly difficult to prototype duration (if the decider is considering a permanent change, then this permanence cannot be prototyped in advance). Another concern is that the prototype, by definition, will not be at the same level of risk because prototyping involves making lower risk decisions in advance of the higher risk decision. Finally, since some modifications can be more easily prototyped than others, this might influence the direction of research on modifications in much the same way that activity trackers that only measure steps might encourage a focus on step-based exercise over other forms of exercise.
When it comes to relying on empirical evidence, one way to strengthen the analogy between the early and late adopters (Stage 2 to Stage 3) is to introduce modifications incrementally. The differences between early and late adopters need not be so strong when the modification in question is a modest change. The worry, though, is that, if we pursue this route, we are deciding not to pursue more dramatic modifications at the moment, and this decision might be the wrong one. We might be able to satisfy preferences about the WIILFA by rolling out modifications quickly.
In this article, in addition to focusing on decisions made about satisfying preferences about the WIILFA, I have been focusing on individual decisions made in isolation. The decisions others make will influence the calculation in a way that complicates things tremendously. For example, those who get value from the novelty factor will find out that the value of their modification decreases as others start to adopt it. For another example, refusing a modification might initially be the right move, but as others adopt it, a variety of pressures might be brought to bear on the refuser that will change the values of the outcomes. Consider the fact that it is increasingly difficult to navigate our modern world without a cell phone, which puts pressure on people to purchase one. Similar types of market and social forces could put pressure on people to modify themselves, which could increase the costs of not modifying. Ultimately, after considering ways to improve on the information we have to inform our own individual decisions, we will need to turn to the herculean task of sorting out rational decisions in a context where the decisions that others make will influence which decisions are rational for us.
1. I am not staking a claim on the nature of well-being by focusing on preference satisfaction, because I am not claiming that satisfying preferences is the key to well-being. Instead, I am simply focusing on the types of decisions that appropriately depend on preference satisfaction. There are many other types of decisions we might make, and some of those other types might be instrumental in improving well-being.
2. Thanks to an anonymous reviewer for this and many other insightful observations.
3. Rachel McKinnon uses the convention “trans*” as an inclusive term referring to a variety of transgender identities.
Agar, Nicholas. 2014. A question about defining moral bioenhancement. Journal of Medical Ethics 40: 369–70.
Archer, Alfred. 2016. Moral enhancement and those left behind. Bioethics 30: 500–510.
Arvan, Marcus. 2015. How to rationally approach life’s transformative experiences. Philosophical Psychology 28: 1199–1218.
Binmore, Ken. 2009. Rational decisions. Princeton, NJ: Princeton University Press.
Bostrom, Nick, and Toby Ord. 2006. The reversal test: Eliminating status quo bias in applied ethics. Ethics 116: 656–79 .
Bostrom, Nick, and Anders Sandberg. 2009. The wisdom of nature: An evolutionary heuristic for human enhancement. In Human enhancement, ed. Julian Savulescu and Nick Bostrom, 375–416. Oxford and New York: Oxford University Press.
Buchak, Lara. 2013. Risk and rationality: Oxford University Press.
Burnett, Bill, and David John Evans. 2016. Designing your life: How to build a well-lived, joyful life. New York: Alfred A. Knopf.
Cabrera, Laura Y. 2011. Memory enhancement: The issues we should not forget about. Journal of Evolution and Technology 22(1): 97–109 .
Chatwin, Caroline, Fiona Measham, Kate O’Brien, and Harry Sumnall. 2017. New drugs, new directions? Research priorities for new psychoactive substances and human enhancement drugs. International Journal of Drug Policy 40: 1–5.
Coeckelbergh, Mark. 2011. Vulnerable cyborgs: Learning to live with our dragons. Journal of Evolution and Technology 22(1): 1–9.
Crittenden, Chris. 2002. Self-deselection: Technopsychotic annihilation via cyborg. Ethics and the Environment 7(2): 127–52.
Daniels, Norman. 2009. Can anyone really be talking about ethically modifying human nature? In Human enhancement, ed. Julian Savulescu and Nick Bostrom, 25–42 . Oxford and NewYork: Oxford University Press.
Dougherty, Tom, Sophie Horowitz, and Paulina Sliwa. 2015. Expecting the unexpected. Res Philosophica 92: 301–21.
Douglas, Thomas. 2015. The harms of enhancement and the conclusive reasons view. Cambridge Quarterly of Healthcare Ethics 24: 23–36.
Dworkin, Ronald. 2002. Sovereign virtue: The theory and practice of equality. Cambridge, MA: Harvard University Press.
Edelman, Shimon. 2018. Identity, immortality, happiness: Pick two. Journal of Evolution and Technology 28(1): 1–17 .
Fukuyama, Francis. 2003. Our posthuman future: Consequences of the biotechnology revolution. New York: Picador.
Gilbert, Daniel T., Matthew A. Killingsworth, Rebecca N. Eyre, and Timothy D. Wilson. 2009. The surprising power of neighborly advice. Science 323:1617– 619.
Graafstra, A. 2013. Biohacking – the forefront of a new kind of human evolution: Amal Graafstra at TEDx SFU. YouTube video file, posted October 17, 2013. Duration 16:38. https://www.youtube.com/watch?time_continue=88&v=7DxVWhFLI6E (accessed May 7, 2019).
Harbisson, Neil, and Moon Ribas. 2018. Cyborg Foundation, history.
https://www.cyborgfoundation.com/ (accessed June 6, 2018).
Hines, Alice. 2018. Magnet implants? Welcome to the world of medical punk. New York Times. May 12.
https://www.nytimes.com/2018/05/12/us/grindfest-magnet-implants-biohacking.html?hp&action=click&pgtype=Homepage&clickSource=story-heading&module=second-column-region®ion=top-news&WT.nav=top-news (accessed May 21, 2018).
Hsee, Christopher K., Reid Hastie, and Jingqiu Chen. 2008. Hedonomics: Bridging decision research with happiness research. Perspectives on Psychological Science 3: 224–43.
Hurwicz, Leonid. 1951. Optimality criteria for decision making under ignorance. Cowles Commission Discussion Paper, Statistics no. 370.
Jeffries, Stuart. 2014. Neil Harbisson: The world’s first cyborg artist. The Guardian. May 6.
https://www.theguardian.com/artanddesign/2014/may/06/neil-harbisson-worlds-first-cyborg-artist (accessed May 21, 2018).
Kahneman, Daniel. 1992. Reference points, anchors, norms, and mixed feelings. Organizational Behavior and Human Decision Processes 51: 296–312.
Krishnamurthy, Meena. 2015. We can make rational decisions to have a child: On the grounds for rejecting L.A. Paul’s arguments. In Permissible progeny?: The morality of procreation and parenting, ed. Sarah Hannan, Samantha Brennan, and Richard Vernon, 170–83. New York: Oxford University Press.
Lara, Francisco. 2017. Oxytocin, empathy and human enhancement. Theoria: An International Journal for Theory, History and Foundations of Science 32: 367–84.
Heath, Chip, Richard Larrick, and George Wu. 1999. Goals as reference points. Cognitive Psychology 38: 79–109.
Liao, S. Matthew, Anders Sandberg, and Rebecca Roache. 2012. Human engineering and climate change. Ethics, Policy & Environment 15: 206–21.
Madrigal, Alexis. 2016. I got a magnet implanted into my finger for science, and it was amazingly weird. Splinter. April 6.
(accessed May 18, 2018).
Martens, Rhonda. 2016. Resource egalitarianism, prosthetics, and enhancements. Journal of Cognition and Neuroethics 3(4): 77–96.
Mather, Mara, and Marcia K Johnson. 2000. Choice-supportive source monitoring: Do our decisions seem better to us as we age? Psychology and Aging 15: 596–606.
Mather, Mara, Marisa Knight, and Michael McCaffrey. 2005. The allure of the alignable: Younger and older adults’ false memories of choice features. Journal of Experimental Psychology: General 134: 38–51.
McKinnon, Rachel. 2015. Trans*formative experiences. Res Philosophica 92: 419–40.
Mihailov, Emilian, and Alexandru Dragomir. 2018. Will cognitive enhancement create post-persons? The use(lessness) of induction in determining the likelihood of moral status enhancement. Bioethics 32: 308–313.
Paul, L. A. 2014. Transformative experience. Oxford: Oxford University Press.
Paul, L. A. 2015a. Transformative choice: Discussion and replies. Res Philosophica 92: 473–545 .
Paul, L. A. 2015b. Transformative experience: Replies to Pettigrew, Barnes and Campbell. Philosophy and Phenomenological Research 91: 794–813.
Paul, L. A. 2015c. What you can’t expect when you’re expecting. Res Philosophica 92: 149–70.
Persson, Ingmar, and Julian Savulescu. 2013. Getting moral enhancement right: The desirability of moral bioenhancement. Bioethics 27: 124–31.
Pettigrew, Richard. 2015. Transformative experience and decision theory.
Philosophy and Phenomenological Research 91: 766–74.
Sample, Ian. 2006. Man rejects first penis implant. The Guardian. September 18.
https://www.theguardian.com/science/2006/sep/18/medicineandhealth.china (accessed May 18, 2018).
Sandel, Michael J. 2009. The case against perfection. Cambridge, MA: Belknap Press.
Savage, Leonard J. 1972. The foundations of statistics. Second ed. New York: Dover.
Skyrms, Brian. 1990. The dynamics of rational deliberation. Cambridge, MA: Harvard
Stix, Madeleine. 2016. World’s first cyborg wants to hack your body. CNN. January 7.
https://www.cnn.com/2014/09/02/tech/innovation/cyborg-neil-harbisson-implant-antenna/index.html (accessed May 9, 2019).
Stoll Benney, Kristen, and Linda A. Henkel. 2006. The role of free choice in memory for past decisions. Memory 14: 1001–1011.
Stoye, Jörg. 2011. Statistical decisions under ambiguity. Theory and Decision 70: 129–48.
Torres, Phil. 2016. Agential risks: A comprehensive introduction. Journal of Evolution and Technology 26(2): 31–47.
Wald, Abraham. 1950. Statistical decision functions. New York: Wiley.
Walsh, Emma, and Peter Ayton. 2009. My imagination versus your feelings: Can personal affective forecasts be improved by knowing other peoples’ emotions? Journal of Experimental Psychology: Applied 15: 351–60.
Weirich, Paul. 2004. Realistic decision theory: Rules for nonideal agents in nonideal circumstances: New York: Oxford University Press.
White, Amy. 2014. Body integrity identity disorder beyond amputation: Consent and liberty. HEC Forum 26: 225–36.
Whitmore, G. A., and M. C. Findlay. 1978. Introduction. In Stochastic dominance: An approach to decision-making under risk, eds G.A. Whitmore and M.C. Findlay, 1–36. N.p.: D.C. Heath.
Wilson, Timothy D., and Daniel T. Gilbert. 2005. Affective forecasting: Knowing what to want. Current Directions in Psychological Science 14: 131–34.